242
complete when each of Canada's 39 distinct Natural Regions was represented by an example that
was considered large enough to sustain most of the
natural processes characteristics of the Natural Region. At the time, 29 of the 39 Natural Regions
were already represented by National Parks, and
there were 10 remaining to be included in the system. The four proposed parks would each be representative of four Natural Regions that were not
yet represented.
This policy context is significant in terms of the
definition of the good. The relevant good that the
average Canadian resident values may be the features of the four specific parks, or the good may be
the contribution that the park makes to the completion of the goal to represent each of the 39 Natural Regions within the National Park System. The
relevant good may be some combination of the two.
If a survey instrument focused exclusively on the
features of the specific parks, outside of the context of the 39 Natural Regions and the fact that the
park in question represented the thirtieth Natural
Region in the system, then the good might be misspecified. It may be that individuals who do not
care to know the details of specific parks do place
a value on the knowledge that remnant examples
of distinct natural regions are preserved within the
system.
Additionally, the fact that there were still ten
Natural Regions to incorporate into the system
might be significant, because the marginal value of
the thirtieth park, in terms of completing the system, may be different from the marginal value of
the park 32 or park 38, or park 39. Economic theory could support either increasing or decreasing
marginal values for additional parks, depending
upon the situation. For example, if completion of
the system generated great value for people, park
39 may be worth more to people than the park 30,
all else equal. On the other hand, the additional
value of creating successive new parks in very remote places, when compared with the demand for
other social programs and public goods, may become less and less, as more of the Natural Regions
are represented. It may be that the specific features
of each park are of less importance than the number of natural regions that are already represented
in the system. If this was the case, then it was feasible that the value of each of the four proposed
parks might depend on the order in which they
would be created. That is, if the marginal value of
each successive park added to the system were declining, then park A may appear to be of greater
value if it was the thirtieth park created, rather than
if it were the thirty-third park created. Since we
could not rule out this possibility, we designed a
Measuring Ecosystem Values
valuation study that asked each respondent to value
one, two, four, or ten parks. The experimental design called for the sample to be split many times
so that the four parks were rotated. The goal was
to estimate the value for each new park in various
combinations of new park creation.
As Rollins and Lyke (1998) demonstrate, it
turned out that the order mattered more than the
specific ecological features of the individual parks.
That is, the demand for new park creation followed
a diminishing marginal value relationship, which
was quite pronounced. Individual Canadian households were willing to pay, on average, about $104,
in terms of a one-time payment, for the thirtieth
park (the first of the four Northwest Territory
parks), regardless of which of the four parks was
created first. The difference between the Natural
Region 30 and Natural Region 31 was about $60
per household, which was statistically significant.
However, the difference between Natural Regions
33 and 39 was not statistically significant. These
results illustrate, first, that the definition of the good
may not be as clear as we would first assume and,
second, that the concept of marginal value is extremely relevant in a policy context. The economic
value of a given ecosystem to society is relative to
the quantity and quality of similar ecosystems that
already exist and relative to other needs that compete for the same financial resources. Park A is
worth relatively less to people if Parks B and C
have already been created, even though they are
different parks. This is one of many reasons why
we may not generally assume that it is possible to
use an ecosystem value that has been estimated in
one context to serve as a proxy for other contexts.
16.6.3 Common Pitfalls
Several of the most common pitfalls that plague
nonmarket valuation of ecosystem goods and services include goods defined in a way that is not relevant for the policy context, attempting to measure
total values when marginal values are needed, confusing average values for marginal values, improper bases for comparison of values, confusion
with regard to the precise definition of the good,
and using a valuation estimate from one site to
serve as a simple proxy for other sites and circumstances. This discussion has touched on several economic concepts that explain some of these
pitfalls. A common reason for most improper uses
of ecosystem valuation is a lack of understanding
of the limitations of the underlying economic theory and methods.
On the other hand, the potential to misuse economic valuation estimates of ecosystems may ex-
complete when each of Canada's 39 distinct Natural Regions was represented by an example that
was considered large enough to sustain most of the
natural processes characteristics of the Natural Region. At the time, 29 of the 39 Natural Regions
were already represented by National Parks, and
there were 10 remaining to be included in the system. The four proposed parks would each be representative of four Natural Regions that were not
yet represented.
This policy context is significant in terms of the
definition of the good. The relevant good that the
average Canadian resident values may be the features of the four specific parks, or the good may be
the contribution that the park makes to the completion of the goal to represent each of the 39 Natural Regions within the National Park System. The
relevant good may be some combination of the two.
If a survey instrument focused exclusively on the
features of the specific parks, outside of the context of the 39 Natural Regions and the fact that the
park in question represented the thirtieth Natural
Region in the system, then the good might be misspecified. It may be that individuals who do not
care to know the details of specific parks do place
a value on the knowledge that remnant examples
of distinct natural regions are preserved within the
system.
Additionally, the fact that there were still ten
Natural Regions to incorporate into the system
might be significant, because the marginal value of
the thirtieth park, in terms of completing the system, may be different from the marginal value of
the park 32 or park 38, or park 39. Economic theory could support either increasing or decreasing
marginal values for additional parks, depending
upon the situation. For example, if completion of
the system generated great value for people, park
39 may be worth more to people than the park 30,
all else equal. On the other hand, the additional
value of creating successive new parks in very remote places, when compared with the demand for
other social programs and public goods, may become less and less, as more of the Natural Regions
are represented. It may be that the specific features
of each park are of less importance than the number of natural regions that are already represented
in the system. If this was the case, then it was feasible that the value of each of the four proposed
parks might depend on the order in which they
would be created. That is, if the marginal value of
each successive park added to the system were declining, then park A may appear to be of greater
value if it was the thirtieth park created, rather than
if it were the thirty-third park created. Since we
could not rule out this possibility, we designed a
Measuring Ecosystem Values
valuation study that asked each respondent to value
one, two, four, or ten parks. The experimental design called for the sample to be split many times
so that the four parks were rotated. The goal was
to estimate the value for each new park in various
combinations of new park creation.
As Rollins and Lyke (1998) demonstrate, it
turned out that the order mattered more than the
specific ecological features of the individual parks.
That is, the demand for new park creation followed
a diminishing marginal value relationship, which
was quite pronounced. Individual Canadian households were willing to pay, on average, about $104,
in terms of a one-time payment, for the thirtieth
park (the first of the four Northwest Territory
parks), regardless of which of the four parks was
created first. The difference between the Natural
Region 30 and Natural Region 31 was about $60
per household, which was statistically significant.
However, the difference between Natural Regions
33 and 39 was not statistically significant. These
results illustrate, first, that the definition of the good
may not be as clear as we would first assume and,
second, that the concept of marginal value is extremely relevant in a policy context. The economic
value of a given ecosystem to society is relative to
the quantity and quality of similar ecosystems that
already exist and relative to other needs that compete for the same financial resources. Park A is
worth relatively less to people if Parks B and C
have already been created, even though they are
different parks. This is one of many reasons why
we may not generally assume that it is possible to
use an ecosystem value that has been estimated in
one context to serve as a proxy for other contexts.
16.6.3 Common Pitfalls
Several of the most common pitfalls that plague
nonmarket valuation of ecosystem goods and services include goods defined in a way that is not relevant for the policy context, attempting to measure
total values when marginal values are needed, confusing average values for marginal values, improper bases for comparison of values, confusion
with regard to the precise definition of the good,
and using a valuation estimate from one site to
serve as a simple proxy for other sites and circumstances. This discussion has touched on several economic concepts that explain some of these
pitfalls. A common reason for most improper uses
of ecosystem valuation is a lack of understanding
of the limitations of the underlying economic theory and methods.
On the other hand, the potential to misuse economic valuation estimates of ecosystems may ex-